Efficient use of quantization parameters in machine-learning models for video coding

1. A method for encoding, by a first encoder, an image block using a first quantization parameter, comprising:

presenting, to an encoder that includes a machine-learning model, the image block and a first value corresponding to the first quantization parameter,

wherein the first value results from a non-linear function using the first quantization parameter as input,

wherein the non-linear function relates to a second function used to calculate, using the first quantization parameter, a Lagrange multiplier that is used in a rate-distortion calculation, and

wherein the machine-learning model is trained to output mode decision parameters for encoding the image block by:

using training data as inputs to the training of the machine-learning model, each training datum including:

a training block that is encoded by a second encoder;

second mode decision parameters used by the second encoder for encoding the training block; and

a second value corresponding to a second quantization parameter,

wherein the second encoder used the second quantization parameter for encoding the training block, and

wherein the second value results from the non-linear function using the second quantization parameter as input;

obtaining first mode decision parameters from the encoder; and

encoding, in a compressed bitstream, the image block using the first mode decision parameters.

2. The method of claim 1 , wherein the image block is a block of an intra-predicted image.

3. The method of claim 2 , wherein the training datum further includes:

first samples of a top neighboring block of the image block; and

second samples of a left neighboring block of the image block.

4. The method of claim 3 , wherein the first mode decision parameters are indicative of a quad-tree partitioning of the image block.

5. The method of claim 1 , wherein the machine-learning model is a convolutional neural-network model.

6. The method of claim 1 , wherein the non-linear function is approximated by linear segments.

7. The method of claim 1 , wherein the non-linear function being an exponential function of the first quantization parameter.

8. The method of claim 7 , wherein the second encoder is an H.264 or an HEVC encoder.

9. The method of claim 1 , wherein the non-linear function being a quadratic function of the first quantization parameter.

10. The method of claim 9 , wherein the second encoder is an H.263, an AV1, or a VP9 encoder.

11. The method of claim 1 , wherein the non-linear function is of a same type as the second function, and the second function is used by the second encoder for determining the Lagrange multiplier used in the rate-distortion calculation.

12. An apparatus for encoding an image block using a first quantization parameter, comprising:

a processor that:

presents, to a machine-learning model, the image block and a first value corresponding to the first quantization parameter,

wherein the first value results from a non-linear function using the first quantization parameter as input,

wherein the non-linear function relates to a second function used to calculate, using the first quantization parameter, a Lagrange multiplier that is used in a rate-distortion calculation, and

wherein the machine-learning model is trained to output mode decision parameters for encoding the image block by:

using training data as inputs to the training of the machine-learning model, each training datum including:

a training block that is encoded by an encoder;

second mode decision parameters used by the encoder for encoding the training block; and

a second value corresponding to a second quantization parameter,

wherein the encoder used the second quantization parameter for encoding the training block, and

wherein the second value results from the non-linear function using the second quantization parameter as input;

obtains first mode decision parameters from the machine-learning model; and

encodes, in a compressed bitstream, the image block using the first mode decision parameters.

13. The apparatus of claim 12 , wherein the image block is a block of an intra-predicted image.

14. The apparatus of claim 13 , wherein the training datum further includes:

first samples of a top neighboring block of the image block; and

second samples of a left neighboring block of the image block.

15. The apparatus of claim 14 , wherein the first mode decision parameters are indicative of a quad-tree partitioning of the image block.

16. The apparatus of claim 12 , wherein the non-linear function is approximated by linear segments.

17. The apparatus of claim 12 , wherein the non-linear function being an exponential function of the first quantization parameter.

18. The apparatus of claim 12 , wherein the non-linear function being a quadratic function of the first quantization parameter.

19. An apparatus for decoding an image block, comprising:

a processor that:

receives, in a compressed bitstream, an indication of a quad-tree partitioning of the image block into sub-blocks, wherein an encoder determined the quad-tree partitioning of the image block using a machine-learning model that is trained by:

using training data as inputs to train the machine-learning model, each training datum including

a training block that is encoded by the encoder,

mode decision parameters used by the encoder for encoding the training block, and

a value corresponding to a quantization parameter,

wherein the encoder used the quantization parameter for encoding the training block,

wherein the value results from a non-linear function using the quantization parameter as input, and

wherein the non-linear function relates to a second function used to calculate, using the quantization parameter, a multiplier that is used in a rate-distortion calculation; and

decodes the image block using the indication of the quad-tree partitioning of the image block.

20. The apparatus of claim 19 , wherein the non-linear function is of a same type as the second function, and the second function is used by the encoder for determining the multiplier used in rate-distortion calculation.